Skip to content

Instantly share code, notes, and snippets.

@universax

universax/tflite_convert.sh

Last active February 16, 2019 09:32
Show Gist options
  • Save universax/2a2f3525e99e3b76dd3dedb630211224 to your computer and use it in GitHub Desktop.
Save universax/2a2f3525e99e3b76dd3dedb630211224 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
tflite_convert.sh
#UINT8
tflite_convert \
--output_file="/home/user_name/output/mobilenet_v1_mymodel_frozen.tflite" \
--graph_def_file="/home/user_name/output/mobilenet_v1_mymodel_frozen.pb" \
--input_arrays=input \
--output_arrays=MobilenetV1/Predictions/Reshape_1 \
--default_ranges_min=0 \
--default_ranges_max=255 \
--inference_type=QUANTIZED_UINT8 \
--mean_values=128 \
--std_dev_values=128
#FLOAT
tflite_convert \
--output_file="/home/user_name/output/mobilenet_v1_mymodel_frozen.tflite" \
--graph_def_file="/home/user_name/output/mobilenet_v1_mymodel_frozen.pb" \
--input_format=TENSORFLOW_GRAPHDEF \
--output_format=TFLITE \
--input_shape=1,224,224,3 \
--input_arrays=input \
--output_arrays=MobilenetV1/Predictions/Reshape_1 \
--inference_type=FLOAT \
--input_data_type=FLOAT
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment